Updated the model and tokenizer to version2, and tagged the current progress to "v2". a28dd1e Amjad-Felemban commited on Jul 14